Research on Spectrum Prediction Technology Based on B-LTF

With the rapid development of global communication technology, problem scarce spectrum resources has become increasingly prominent. In order to alleviate frequency use, rationally use limited and improve utilization, prediction technology emerged. Through effective usage, number subsequent sensing processes can be slowed down, accuracy decisions increased response speed whole cognitive radio technology. The rise deep learning brought changes traditional predicting algorithms. This paper proposes a method called Back Propagation-Long short-term memory Time Forecasting (B-LTF) by using Short-term Memory (BP-LSTM) network model. According historical data, future trend channel state time node are predicted. purpose our research is achieve dynamic access improving better assisting By comparing with BP, LSTM Gate Recurrent Unit (GRU) models, we clarify that improved model recurrent deal series more effectively. simulation results show proposed performance, change in length significant impact on